Computer-implemented method of browsing a resource through voice interaction comprising the following steps: A. acquiring (100) from a user a request aimed at browsing a resource; B. downloading (130) the requested resource; C. performing a syntactic parsing (135) of the downloaded resource; D. extracting (150) from the downloaded resource one or more lists, if any, of selectable shortcuts pointing to portions inside or outside the downloaded resource through a syntactic analysis and/or a semantic analysis and/or a morphological-visual analysis of extraction of lists of selectable shortcuts on the basis of an ontology (245) corresponding to the type of resource; E. on the basis of the ontology (245) corresponding to the type of resource, building (225) a list of one or more lists of selectable shortcuts extracted in step D ordered according to a list prioritisation; F. extracting (150) from the downloaded resource one or more content elements through a syntactic analysis and/or a semantic analysis and/or a morphological-visual analysis of extraction of content elements on the basis of the ontology (245) corresponding to the type of resource; G. on the basis of the ontology (245) corresponding to the type of resource, building (290) a list of content elements extracted in step F ordered according to a content element prioritisation; H. on the basis of the lists built in steps E and G and on the basis of the ontology (245) corresponding to the type of resource, building a final structure of lists of selectable shortcuts and of content elements; I. playing (125) a voice prompt based on the final structure and starting a voice interaction with the user for browsing the resource.

Computer-implemented method of browsing a resource through voice interaction comprising the following steps: A. acquiring (100) from a user a request aimed at browsing a resource; B. downloading (130) the requested resource; C. performing a syntactic parsing (135) of the downloaded resource; D. extracting (150) from the downloaded resource one or more lists, if any, of selectable shortcuts pointing to portions inside or outside the downloaded resource through a syntactic analysis and/or a semantic analysis and/or a morphological-visual analysis of extraction of lists of selectable shortcuts on the basis of an ontology (245) corresponding to the type of resource; E. on the basis of the ontology (245) corresponding to the type of resource, building (225) a list of one or more lists of selectable shortcuts extracted in step D ordered according to a list prioritisation; F. extracting (150) from the downloaded resource one or more content elements through a syntactic analysis and/or a semantic analysis and/or a morphological-visual analysis of extraction of content elements on the basis of the ontology (245) corresponding to the type of resource; G. on the basis of the ontology (245) corresponding to the type of resource, building (290) a list of content elements extracted in step F ordered according to a content element prioritisation; H. on the basis of the lists built in steps E and G and on the basis of the ontology (245) corresponding to the type of resource, building a final structure of lists of selectable shortcuts and of content elements; I. playing (125) a voice prompt based on the final structure and starting a voice interaction with the user for browsing the resource.

The system receives, from a client, a first request for a document which is parsed based on a syntax. The system parses and returns an initial number of lines of the document, thereby allowing the client to display the parsed initial number of lines in a visible window. Subsequent to successfully parsing a remainder of the lines, the system stores a fully-parsed version. Responsive to a scrolling action in the visible window, the system provides a relevant portion of an unparsed document associated with the first request. Responsive to detecting a termination of the scrolling action, the system receives a second request for parsed lines corresponding to a first set of lines. The system returns the corresponding parsed lines, thereby allowing the client to display the corresponding parsed lines in the visible window.

The system receives, from a client, a first request for a document which is parsed based on a syntax. The system parses and returns an initial number of lines of the document, thereby allowing the client to display the parsed initial number of lines in a visible window. Subsequent to successfully parsing a remainder of the lines, the system stores a fully-parsed version. Responsive to a scrolling action in the visible window, the system provides a relevant portion of an unparsed document associated with the first request. Responsive to detecting a termination of the scrolling action, the system receives a second request for parsed lines corresponding to a first set of lines. The system returns the corresponding parsed lines, thereby allowing the client to display the corresponding parsed lines in the visible window.

Disclosed are a method for training a semantic representation model, a device and a storage medium, which relate to the field of computer technologies, and particularly to the field of artificial intelligence, such as a natural language processing technology, a deep learning technology, or the like. The method for training a semantic representation model includes: obtaining an anchor sample based on a sentence, and obtaining a positive sample and a negative sample based on syntactic information of the sentence; processing the anchor sample, the positive sample and the negative sample using the semantic representation model respectively, so as to obtain an anchor-sample semantic representation, a positive-sample semantic representation and a negative-sample semantic representation; constructing a contrast loss function based on the anchor-sample semantic representation, the positive-sample semantic representation, and the negative-sample semantic representation; and training the semantic representation model based on the contrast loss function.

Disclosed are a method for training a semantic representation model, a device and a storage medium, which relate to the field of computer technologies, and particularly to the field of artificial intelligence, such as a natural language processing technology, a deep learning technology, or the like. The method for training a semantic representation model includes: obtaining an anchor sample based on a sentence, and obtaining a positive sample and a negative sample based on syntactic information of the sentence; processing the anchor sample, the positive sample and the negative sample using the semantic representation model respectively, so as to obtain an anchor-sample semantic representation, a positive-sample semantic representation and a negative-sample semantic representation; constructing a contrast loss function based on the anchor-sample semantic representation, the positive-sample semantic representation, and the negative-sample semantic representation; and training the semantic representation model based on the contrast loss function.

Methods and apparatus for automated processing of natural language text is described. The text can be preprocessed to produce language-space data that includes descriptive data elements for words. Source code that includes linguistic expressions, and that may be written in a programming language that is user-friendly to linguists, can be compiled to produce finite-state transducers and bi-machine transducers that may be applied directly to the language-space data by a language-processing virtual machine. The language-processing virtual machine can select and execute code segments identified in the finite-state and/or bi-machine transducers to disambiguate meanings of words in the text.

Methods and apparatus for automated processing of natural language text is described. The text can be preprocessed to produce language-space data that includes descriptive data elements for words. Source code that includes linguistic expressions, and that may be written in a programming language that is user-friendly to linguists, can be compiled to produce finite-state transducers and bi-machine transducers that may be applied directly to the language-space data by a language-processing virtual machine. The language-processing virtual machine can select and execute code segments identified in the finite-state and/or bi-machine transducers to disambiguate meanings of words in the text.

Computer-implemented method of browsing a resource through voice interaction comprising the following steps: A. acquiring (100) from a user a request aimed at browsing a resource; B. downloading (130) the requested resource; C. performing a syntactic parsing (135) of the downloaded resource; D. extracting (150) from the downloaded resource one or more lists, if any, of selectable shortcuts pointing to portions inside or outside the downloaded resource through a syntactic analysis and/or a semantic analysis and/or a morphological-visual analysis of extraction of lists of selectable shortcuts on the basis of an ontology (245) corresponding to the type of resource; E. on the basis of the ontology (245) corresponding to the type of resource, building (225) a list of one or more lists of selectable shortcuts extracted in step D ordered according to a list prioritisation; F. extracting (150) from the downloaded resource one or more content elements through a syntactic analysis and/or a semantic analysis and/or a morphological-visual analysis of extraction of content elements on the basis of the ontology (245) corresponding to the type of resource; G. on the basis of the ontology (245) corresponding to the type of resource, building (290) a list of content elements extracted in step F ordered according to a content element prioritisation; H. on the basis of the lists built in steps E and G and on the basis of the ontology (245) corresponding to the type of resource, building a final structure of lists of selectable shortcuts and of content elements; I. playing (125) a voice prompt based on the final structure and starting a voice interaction with the user for browsing the resource.

Computer-implemented method of browsing a resource through voice interaction comprising the following steps: A. acquiring (100) from a user a request aimed at browsing a resource; B. downloading (130) the requested resource; C. performing a syntactic parsing (135) of the downloaded resource; D. extracting (150) from the downloaded resource one or more lists, if any, of selectable shortcuts pointing to portions inside or outside the downloaded resource through a syntactic analysis and/or a semantic analysis and/or a morphological-visual analysis of extraction of lists of selectable shortcuts on the basis of an ontology (245) corresponding to the type of resource; E. on the basis of the ontology (245) corresponding to the type of resource, building (225) a list of one or more lists of selectable shortcuts extracted in step D ordered according to a list prioritisation; F. extracting (150) from the downloaded resource one or more content elements through a syntactic analysis and/or a semantic analysis and/or a morphological-visual analysis of extraction of content elements on the basis of the ontology (245) corresponding to the type of resource; G. on the basis of the ontology (245) corresponding to the type of resource, building (290) a list of content elements extracted in step F ordered according to a content element prioritisation; H. on the basis of the lists built in steps E and G and on the basis of the ontology (245) corresponding to the type of resource, building a final structure of lists of selectable shortcuts and of content elements; I. playing (125) a voice prompt based on the final structure and starting a voice interaction with the user for browsing the resource.